IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 20 March 2008 (all times are UTC)
| 00:02 | MacIver has quit IRC | |
| 00:07 | <johnny> hmm..
| |
| 00:07 | vagrantc, you about?
| |
| 00:09 | <vagrantc> johnny: yes, but i'd best be getting not about
| |
| 00:09 | <johnny> ?
| |
| 00:09 | leaving?
| |
| 00:09 | <vagrantc> yeah ... just dropped in to send a couple emails ... and habitually fired up irc :)
| |
| 00:09 | <johnny> oh.. sorry
| |
| 00:10 | <vagrantc> something quick ?
| |
| 00:10 | <johnny> i just wondered where i could see the kernel commandline you pass to the initrd
| |
| 00:10 | <vagrantc> johnny: egrep -r BOOTPROMPT_OPTS
| |
| 00:10 | in the ltsp sources
| |
| 00:10 | <johnny> oh.. aha.. i was trying to examine the intramfs..
| |
| 00:19 | MacIver has joined #ltsp | |
| 00:26 | abadger1999 has quit IRC | |
| 00:26 | tux_440volt has quit IRC | |
| 00:26 | oh207 has quit IRC | |
| 00:26 | Topslack has quit IRC | |
| 00:26 | lns has quit IRC | |
| 00:26 | otavio has quit IRC | |
| 00:26 | Shingoshi has quit IRC | |
| 00:26 | BadMagic has quit IRC | |
| 00:26 | moquist has quit IRC | |
| 00:26 | gu_ has quit IRC | |
| 00:26 | bjorn has quit IRC | |
| 00:27 | BadMagic has joined #ltsp | |
| 00:29 | sbalneav has quit IRC | |
| 00:29 | rjune_ has quit IRC | |
| 00:29 | F-GT has quit IRC | |
| 00:29 | Patina has quit IRC | |
| 00:29 | vagrantc has quit IRC | |
| 00:30 | rjune_ has joined #ltsp | |
| 00:30 | abadger1999 has joined #ltsp | |
| 00:30 | tux_440volt has joined #ltsp | |
| 00:30 | oh207 has joined #ltsp | |
| 00:30 | Topslack has joined #ltsp | |
| 00:30 | lns has joined #ltsp | |
| 00:30 | otavio has joined #ltsp | |
| 00:30 | Shingoshi has joined #ltsp | |
| 00:30 | moquist has joined #ltsp | |
| 00:30 | bjorn has joined #ltsp | |
| 00:30 | gu_ has joined #ltsp | |
| 00:30 | otavio has quit IRC | |
| 00:31 | sbalneav has joined #ltsp | |
| 00:31 | F-GT has joined #ltsp | |
| 00:31 | Patina has joined #ltsp | |
| 00:33 | tux_440volt has quit IRC | |
| 01:01 | indradg__ has joined #ltsp | |
| 01:02 | oh207 has quit IRC | |
| 01:03 | oh207 has joined #ltsp | |
| 01:04 | oh207 has joined #ltsp | |
| 01:05 | indradg has joined #ltsp | |
| 01:14 | indradg__ has quit IRC | |
| 01:19 | indradg_ has quit IRC | |
| 01:37 | basanta has quit IRC | |
| 01:37 | basanta_ has joined #ltsp | |
| 02:08 | indradg_ has joined #ltsp | |
| 02:25 | indradg has quit IRC | |
| 02:45 | Pascal_1 has joined #ltsp | |
| 02:58 | <Pascal_1> hello
| |
| 02:58 | what is the way to know the version of ldm installed on the server ltsp , i've got to report a bug .is it the ldm in the chroot environment ?
| |
| 03:00 | daya has joined #ltsp | |
| 03:00 | <Pascal_1> when i do a apt-cache show ldm
| |
| 03:01 | in the chroot i've got 2 version of ldm and i dont know wich is installed
| |
| 03:13 | makghosh has joined #ltsp | |
| 03:26 | <Pascal_1> i made a reportbug but i'm not sure it's clear. someone coulmd told me if it is ? : http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=471793
| |
| 03:28 | deavid has joined #ltsp | |
| 03:46 | tux_440volt has joined #ltsp | |
| 03:47 | <daduke> Pascal_1: dpkg --root=/opt/ltsp/i386 -l rdesktop | egrep ^ii show you the installed packages incl. version
| |
| 03:48 | <Pascal_1> hello daduke !! too late ;-) i made the reportbug vagran told me it was a bug
| |
| 03:49 | <daduke> tut
| |
| 03:49 | <Pascal_1> tut ?
| |
| 03:49 | and the problem was the same with the two version
| |
| 03:49 | <daduke> wrong window focus, sorry ;(
| |
| 03:49 | ah, ok. so it is a bug after all.
| |
| 03:51 | <Pascal_1> yes i'm sad !!
| |
| 03:51 | <daduke> why sad? it's a bug, it will be fixed by vagrantc, and everything will be fine.
| |
| 03:53 | <Pascal_1> ok
| |
| 03:53 | then i'm happy !!! ;-)
| |
| 03:58 | i'm in a hurry
| |
| 03:58 | ltsp is a very good solution
| |
| 04:00 | <daduke> indeed
| |
| 04:53 | makghosh has quit IRC | |
| 04:54 | tux_440volt has quit IRC | |
| 05:23 | vlt has quit IRC | |
| 05:33 | daya has quit IRC | |
| 05:36 | Faithful has quit IRC | |
| 05:47 | xcasex has quit IRC | |
| 06:03 | ogra has quit IRC | |
| 06:03 | ogra has joined #ltsp | |
| 06:04 | DonSilver has joined #ltsp | |
| 06:16 | ogra has quit IRC | |
| 06:16 | ogra__ has quit IRC | |
| 06:16 | ogra has joined #ltsp | |
| 06:17 | ogra__ has joined #ltsp | |
| 06:34 | subir has quit IRC | |
| 06:52 | basanta_ has quit IRC | |
| 06:56 | Guaraldo has joined #ltsp | |
| 07:00 | xcasex has joined #ltsp | |
| 07:17 | cliebow_ has joined #ltsp | |
| 07:19 | xcasex_ has joined #ltsp | |
| 07:27 | xcasex has quit IRC | |
| 07:42 | ogra has quit IRC | |
| 07:43 | slidesinger has joined #ltsp | |
| 07:51 | ElDios has joined #ltsp | |
| 07:51 | <ElDios> hello everyone
| |
| 07:51 | simple question (perhaps)
| |
| 07:51 | is it possible to estimate the average consume per LTSP client ?
| |
| 07:52 | (with basic features)
| |
| 07:52 | bandwidth consume*
| |
| 07:53 | <tarzeau> ElDios: hm... let's see with iptraf
| |
| 07:53 | <ElDios> thnx ^_^ tarzeau
| |
| 07:54 | <tarzeau> ElDios: we've got like 1,5mb/s on our server with 20 clients out... let's see
| |
| 07:55 | <ElDios> thnx.. meanwhile I'll BRB
| |
| 07:57 | DonSilver has quit IRC | |
| 08:03 | tux_440volt has joined #ltsp | |
| 08:06 | Faithful has joined #ltsp | |
| 08:08 | xcasex_ is now known as xcasex | |
| 08:10 | <Pascal_1> hello, i've got an other problem i put some action in the .bash_logout of a user, when the user logout from thin client those action are not launched. anybody knows about this ?
| |
| 08:12 | mhterres has joined #ltsp | |
| 08:12 | <laga> Pascal_1: .bash_logout is probably not used because your users dont use bash..
| |
| 08:13 | <ElDios> Pascal_1 how do you quit the session?
| |
| 08:13 | <Pascal_1> i dont really know about the ldm process
| |
| 08:13 | ElDios, in french "clore la session"
| |
| 08:13 | close session
| |
| 08:14 | when a user loggin from a thin client (ldm) and he logout the bash_logout is not read ?
| |
| 08:14 | <ElDios> it is read when you *close* (nor kill or anything) a bash session
| |
| 08:15 | you should define to run .bash_logout when the session is closed
| |
| 08:15 | the ltsp session is closed :)
| |
| 08:16 | tarzeau any results?
| |
| 08:16 | <tarzeau> ElDios: well 14 users logged in, 1,5mb, about 100kb/client
| |
| 08:17 | ElDios: but when the soccer games start, they'll all be watching it, i guess it'll need about 1mb/client then
| |
| 08:17 | but even that is no problem, since our server has gbit, and the clients 100mbit (making 10mb/s)
| |
| 08:18 | the cpu load is low too, as well as the memory usage
| |
| 08:18 | <ElDios> 110kbit/client is not that low
| |
| 08:18 | <tarzeau> no?
| |
| 08:18 | <ElDios> I mean.. think about using it over VPN!
| |
| 08:18 | <tarzeau> kbyte, not kbit
| |
| 08:19 | where do you live?
| |
| 08:19 | <ElDios> even worse :)
| |
| 08:19 | near Milan
| |
| 08:19 | <tarzeau> istanbul? anal internet?
| |
| 08:19 | <ElDios> :)
| |
| 08:19 | ahhh
| |
| 08:19 | ahah
| |
| 08:19 | nono that's not so bad
| |
| 08:19 | <tarzeau> heck even they got the cableshitinternet now
| |
| 08:19 | <ElDios> the problem isn't the download band
| |
| 08:19 | <tarzeau> yes it used to be, when you run traceroute to www.superonline.com.tr
| |
| 08:19 | <ElDios> it's the upload side
| |
| 08:19 | <tarzeau> you either had timeouts or minimum of 500ms
| |
| 08:20 | how's that the problem? i was giving you the details of input/output, not just download or just upload
| |
| 08:20 | the clients usually download more than they upload
| |
| 08:20 | since the clients are the displays
| |
| 08:20 | and they show stuff
| |
| 08:20 | and X sucks, since it's not display postscript
| |
| 08:20 | <ElDios> yep
| |
| 08:20 | <tarzeau> daduke: do we use compressed remote x sessions or not?
| |
| 08:20 | <ElDios> so the server must have an high bandwidth
| |
| 08:21 | <tarzeau> ElDios: it all depends what the clients exactly do
| |
| 08:21 | <ElDios> mainly office work
| |
| 08:21 | <tarzeau> some x software sucks horrbily, if you just run some textmode programs in some terminals... it's all different
| |
| 08:22 | well firefox with animated web and openoffice suck for these things
| |
| 08:22 | can you define office work? here that's writing latex documents in something like emacs
| |
| 08:22 | <daduke> tarzeau: we use an ssh tunnel for X traffic
| |
| 08:22 | <tarzeau> which works very good on those clients
| |
| 08:22 | daduke: with or without -C ?
| |
| 08:22 | <ElDios> tarzeau openoffice, and firefox should be fine
| |
| 08:23 | <tarzeau> no firefox sucks if used remotely
| |
| 08:24 | actually, firefox sucks. (no matter how)
| |
| 08:25 | <ElDios> ahah :)
| |
| 08:25 | so what graphical browser would you suggest?
| |
| 08:26 | <daduke> tarzeau: w/o.
| |
| 08:26 | <tarzeau> personally? i use links2 -g and the old thing on my old openstep box
| |
| 08:26 | ElDios: http://gnu.ethz.ch/www.levenez.com/theinternet.jpg
| |
| 08:26 | <daduke> tarzeau: links2 has what? -4234 points in acid3?
| |
| 08:27 | <tarzeau> i don't care how shabby it looks, i care about the graphics and letters
| |
| 08:27 | <ElDios> mmm.. oke
| |
| 08:27 | <tarzeau> there's a saying:
| |
| 08:27 | "One could say that GNUstep is a very nice woman, but without proper make-up and with a shabby dress. Although the expert eye could see the star sparkle, the average person maybe would prefer a more normal female with choosen make-up and dress."
| |
| 08:28 | <ElDios> what about eclipse? anyone ever tried?
| |
| 08:28 | <tarzeau> eclipse? the people who tried it here switched back to emacs
| |
| 08:28 | (the ones that i know, since i had to package that crap for them, since it was too hard for them to install themselves)
| |
| 08:28 | <daduke> tarzeau: it's nice for Java development
| |
| 08:28 | <ElDios> I was speaking again of the bandwidth consume
| |
| 08:28 | <tarzeau> i'm glad i don't need to do java development then
| |
| 08:28 | <ElDios> ^_^
| |
| 08:30 | tux_440volt has quit IRC | |
| 08:34 | Gadi has joined #ltsp | |
| 08:34 | * warren thought of a way around needing DNS | |
| 08:34 | * warren silent.wav! | |
| 08:34 | <Gadi> lol
| |
| 08:35 | * Gadi missed the joke but still laughs at the punchline :) | |
| 08:36 | dtrask has joined #ltsp | |
| 08:39 | abadger1999 has quit IRC | |
| 08:42 | mccann has joined #ltsp | |
| 08:58 | ogra_cmpc has joined #ltsp | |
| 09:05 | abadger1999 has joined #ltsp | |
| 09:18 | DonSilver has joined #ltsp | |
| 09:18 | deavid has quit IRC | |
| 09:30 | DonSilver has quit IRC | |
| 09:31 | DonSilver has joined #ltsp | |
| 09:35 | Faithful has quit IRC | |
| 09:52 | * warren pushes 666 | |
| 09:53 | <cyberorg> the evil one?
| |
| 09:53 | makghosh has joined #ltsp | |
| 10:01 | <Pascal_1> is there a way to execute command when user disconnect from ldm ?
| |
| 10:01 | <warren> like whta?
| |
| 10:02 | DonSilver has quit IRC | |
| 10:03 | DonSilver has joined #ltsp | |
| 10:03 | <Pascal_1> warren, ??
| |
| 10:05 | <mnemoc> Pascal_1: what do you want to execute?
| |
| 10:05 | <Pascal_1> hmm for example my problem to umount samba share at logout
| |
| 10:06 | actually it seems there is a bug concerning ldm/pam which make that i cant do it with pam_mount
| |
| 10:06 | mnemoc, it's an example
| |
| 10:11 | <mnemoc> Pascal_1: i was just translating warren's question :-) wait for his comments :)
| |
| 10:14 | DonSilver has quit IRC | |
| 10:15 | <warren> hmm
| |
| 10:15 | <ogra_cmpc> hmm?
| |
| 10:15 | <warren> The latest changes to ldm break its build
| |
| 10:15 | gcc -DHAVE_CONFIG_H -I. -I.. -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include -DRC_DIR=\""/usr/share/ldm"\" -DLDM_EXEC_DIR=\""/usr/libexec/ldm"\" -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic -MT sshutils.o -MD -MP -MF .deps/sshutils.Tpo -c -o sshutils.o sshutils.c
| |
| 10:15 | ldm-getenv-bool.c: In function 'ldm_getenv_bool':
| |
| 10:15 | ldm-getenv-bool.c:16: error: 'TRUE' undeclared (first use in this function)
| |
| 10:16 | <mnemoc> why not using stdbool.h?
| |
| 10:16 | <warren> a bunch of includes were removed
| |
| 10:17 | anyone know what version of gcc vagrant is using?
| |
| 10:17 | <mnemoc> i mean, stdbool.h exports true, not TRUE
| |
| 10:17 | <warren> this might be a bug exposed by gcc's reduced includes
| |
| 10:17 | gcc-4.3 that is
| |
| 10:17 | vagrantc has joined #ltsp | |
| 10:17 | <ogra_cmpc> i'm having 4.2.3-2ubuntu1
| |
| 10:18 | so its unlikely vagrantc uses something newer
| |
| 10:18 | <vagrantc> ??
| |
| 10:18 | <ogra_cmpc> speaking of the devil
| |
| 10:18 | mccann has quit IRC | |
| 10:18 | <ogra_cmpc> vagrantc, gcc
| |
| 10:18 | <warren> vagrantc, what version of gcc are you using?
| |
| 10:18 | vagrantc, you broke ldm build! =)
| |
| 10:19 | <Pascal_1> hello vagrantc !
| |
| 10:19 | have you seen my reportbug? is it clear ?
| |
| 10:20 | <ltsppbot> Someone pasted "Executing(%prep): /bin/sh -e /" (173 lines) at http://pastebot.ltsp.org/480
| |
| 10:20 | <warren> vagrantc, http://pastebot.ltsp.org/480
| |
| 10:20 | vagrantc, I'm wondering what version of gcc you are using because it matters in this case.
| |
| 10:20 | mccann has joined #ltsp | |
| 10:21 | <ogra_cmpc> i dont think it matters, to that looks like there was one header to much dropped
| |
| 10:21 | s/to/to me
| |
| 10:22 | <warren> I'm assuming he actually tested it before checking in? =)(
| |
| 10:22 | <Pascal_1> warren, any idea for my question ?
| |
| 10:22 | <warren> Pascal_1, sorry no
| |
| 10:24 | <vagrantc> warren: well, i built it with a version of gcc that was recently in debian unstable ...
| |
| 10:24 | but it's recently transitioned
| |
| 10:24 | <warren> any idea what include is missing?
| |
| 10:24 | <ogra_cmpc> the one that gives you TRUE
| |
| 10:25 | <vagrantc> i applied the patch from scottie's branch for the headers
| |
| 10:25 | ah!
| |
| 10:25 | i didn't apply that part of the patch
| |
| 10:25 | <ogra_cmpc> TRUE is used in ldm.c
| |
| 10:25 | so it must be one of these
| |
| 10:25 | <vagrantc> TRUE is used in ldm_getenv_bool.c
| |
| 10:25 | * warren waits | |
| 10:25 | <ogra_cmpc> vagrantc, yes, thats where it fails :)
| |
| 10:25 | <vagrantc> scotty had switched it from TRUE -> 1 and FALSE -> 0 ... but i guess i didn't test that
| |
| 10:25 | <warren> vagrantc, did you test scottie's paste for the xauth fix?
| |
| 10:26 | <vagrantc> warren: no, haven't had a chance
| |
| 10:26 | <ogra_cmpc> vagrantc, the other code uses TRUE without prob
| |
| 10:26 | so there must be a header thet introduces it
| |
| 10:26 | <vagrantc> hope so
| |
| 10:26 | <ogra_cmpc> we should rather stay consistent
| |
| 10:26 | <vagrantc> as TRUE FALSE is much clearer than 0 1
| |
| 10:27 | especially when switching between shell and C
| |
| 10:27 | worst case is patch it to use 1 or 0
| |
| 10:27 | * ogra_cmpc tries to remember the name of the package containing the header manpages | |
| 10:28 | <ogra_cmpc> ah, manpages-dev
| |
| 10:28 | <vagrantc> ok. really. i won't check things in without testing them anymore. i promise.
| |
| 10:28 | cesar_bo has joined #ltsp | |
| 10:29 | <vagrantc> ogra_cmpc: i'll be sharing a continent with you for the first week of april ...
| |
| 10:30 | <ogra_cmpc> oh, nice, where do you go ?
| |
| 10:31 | <vagrantc> ogra_cmpc: extremadura, spain
| |
| 10:31 | <ogra_cmpc> ah, nice
| |
| 10:33 | cesar_bo has quit IRC | |
| 10:33 | slidesinger has quit IRC | |
| 10:34 | <vagrantc> Pascal_1: yes, your bug report is good enough.
| |
| 10:34 | <Pascal_1> ok thanks
| |
| 10:34 | cesar_bo has joined #ltsp | |
| 10:34 | <vagrantc> Pascal_1: would be nice to get a snippet from your auth log of what a normal ssh session looks like and what an ldm login looks like.
| |
| 10:34 | <Pascal_1> hmmhow i do that ?
| |
| 10:34 | mail on the bug to add
| |
| 10:34 | ?
| |
| 10:35 | <vagrantc> Pascal_1: email 471793@bugs.debian.org
| |
| 10:35 | <Pascal_1> in the body i put only the logs ? nothing in subject ?
| |
| 10:36 | <ogra_cmpc> vagrantc, you need types.h i think
| |
| 10:36 | <cesar_bo> Hi all, I had a ltsp network running pretty well, I want to enable a printer for a few clients, and fallow the documentation, adding the lines PRINTER_0_DEVICE=/dev/lp0, on the default head on the lts.conf file of the client /opt/ltsp/i386, but it doesn't work.
| |
| 10:36 | <vagrantc> Pascal_1: well, put something meaningful in the subject and body just like you would a normal email ...
| |
| 10:37 | slidesinger has joined #ltsp | |
| 10:37 | <vagrantc> ogra_cmpc: well, we all need types.h, then :)
| |
| 10:37 | <Pascal_1> i put the logs in the body
| |
| 10:37 | ?
| |
| 10:37 | <ogra_cmpc> vagrantc, wasnt my patch :P
| |
| 10:38 | <vagrantc> Pascal_1: just cut the part of the logs and comment on them ... like ... this is what a normal ssh login looks like: ... lines from auth.log ... this is what an ldm login looks like: ... lines from auth.log ...
| |
| 10:39 | ogra_cmpc: wasn't my fix, either :)
| |
| 10:39 | ogra_cmpc: but if you'd rather i test and commit it, i will.
| |
| 10:40 | one of these days, i'm going to do a bug reporting class at freegeek :)
| |
| 10:40 | <ogra_cmpc> vagrantc, i dont even have that new code here yet
| |
| 10:41 | <vagrantc> ogra_cmpc: i'll test it ...
| |
| 10:42 | <warren> vagrantc, so wait, did you find a fix for this?
| |
| 10:42 | cesar_bo has left #ltsp | |
| 10:42 | <vagrantc> warren: ogra_cmpc suggested to add types.h
| |
| 10:42 | <ogra_cmpc> sys/types.h actually
| |
| 10:43 | <vagrantc> warren: i can try and build a package to test
| |
| 10:43 | warren: unless you want to ...
| |
| 10:43 | <warren> I can do it
| |
| 10:43 | which is the preferred file to add the include?
| |
| 10:43 | <vagrantc> or we can race :)
| |
| 10:44 | <ogra_cmpc> warren, ldm-getenv-bool.c needs to know about TRUE and FALSE
| |
| 10:44 | so add it there
| |
| 10:47 | <warren> Has anyone tested scottie's paste for the xauth issue?
| |
| 10:47 | <vagrantc> i *think* it's dependent on other things being functional
| |
| 10:47 | but i haven't tested it
| |
| 10:47 | <warren> like what?
| |
| 10:47 | <vagrantc> like ldm generating the xauth files
| |
| 10:47 | <warren> X called with our options creates /root/.Xauthority ?
| |
| 10:48 | vagrantc, his script isn't one of the rc.d scripts that happens after ldm is running?
| |
| 10:48 | <vagrantc> warren: it is one of the rc.d scripts
| |
| 10:48 | * warren suspects that it will just work | |
| 10:48 | <vagrantc> /root/.Xauthority is produced, but i don't think it contains anything
| |
| 10:48 | * warren tries it | |
| 10:49 | <warren> does it need to contain anything?
| |
| 10:49 | <vagrantc> looking at scotties patch, i think so
| |
| 10:49 | if you want to just go ahead and test it rather than question my thoughts on the matter, please do.
| |
| 10:49 | <warren> I'll figure out what's needed today
| |
| 10:49 | <vagrantc> i could be wrong, but i figured i'd give a heads up
| |
| 10:49 | <warren> you are likely right
| |
| 10:50 | <vagrantc> it also may be the case that it works for LDM_DIRECTX=true, but breaks LDM_DIRECTX=false
| |
| 10:50 | <warren> heh
| |
| 10:50 | btw where am I supposed to define LDM_DIRECTX=true?
| |
| 10:51 | * vagrantc wonders what warren has been smoking | |
| 10:51 | <vagrantc> :)
| |
| 10:51 | same place we define most all ltsp configuration options ... lts.conf
| |
| 10:51 | <warren> so like, within the [default] block?
| |
| 10:52 | <vagrantc> that will enable it for all thin clients ...
| |
| 10:52 | and within a client-specific block will enable it for ... that client
| |
| 10:55 | Patina_ has joined #ltsp | |
| 10:56 | <warren> thanks
| |
| 10:57 | ogra_cmpc, ldm-getenv-bool.c alone wasn't good enough
| |
| 10:57 | <ogra_cmpc> oh
| |
| 10:59 | warren,
| |
| 10:59 | #define TRUE 1
| |
| 10:59 | #define FALSE 0
| |
| 10:59 | put that at the top
| |
| 11:00 | * mnemoc wonders why don't you use POSIX's stdbool.h | |
| 11:00 | <vagrantc> warren: are you saying you still got an undefined error for TRUE/FALSE, or you encountered additional errors?
| |
| 11:01 | mnemoc: patches accepted
| |
| 11:01 | presuming, of course, they improve things :)
| |
| 11:01 | <ogra_cmpc> mnemoc, because all the rest of the code uses TRUE, not true
| |
| 11:01 | mnemoc, nobody will refuse a patch that fixes it everywhere though
| |
| 11:02 | <mnemoc> ogra_cmpc: ok, i'll try to send it today
| |
| 11:03 | <vagrantc> warren: definitely fixes the TRUE/FALSE build error.
| |
| 11:03 | warren: but another patch of mine appears to be breaking.
| |
| 11:04 | warren: well, at least fixes it for me.
| |
| 11:05 | or maybe not.
| |
| 11:05 | i don't get the TRUE/FALSE warnings even without adding sys/types.h
| |
| 11:06 | <ogra_cmpc> weird
| |
| 11:06 | <vagrantc> i do get: ldm.c:477: error: too few arguments to function 'launch_x'
| |
| 11:06 | which is clearly from my patches
| |
| 11:06 | as i added arguments to launch_x ...
| |
| 11:07 | but i tested those patches before ...
| |
| 11:08 | i am very confused.
| |
| 11:09 | i must have screwed up in the merge, as it's present in my original branch
| |
| 11:09 | warren: using gcc 4.2.3 today
| |
| 11:10 | mccann has quit IRC | |
| 11:13 | <vagrantc> ah, now i get TRUE/FALSE errors.
| |
| 11:13 | sys/types.h doesn't fix it ...
| |
| 11:13 | <ogra_cmpc> put a define line in for each ... mapping them to 0 and 1
| |
| 11:14 | and forget about the header
| |
| 11:14 | <vagrantc> define TRUE 1
| |
| 11:14 | ??
| |
| 11:14 | <ogra_cmpc> with a hash in front
| |
| 11:14 | #define TRUE 1
| |
| 11:14 | #define FALSE 0
| |
| 11:15 | <vagrantc> looks promising
| |
| 11:15 | Pascal_1 has quit IRC | |
| 11:16 | <vagrantc> ok, it builds ...
| |
| 11:17 | could just revert the patch wholesale.
| |
| 11:17 | we'll get it when scotty fixes everything :)
| |
| 11:19 | <ogra_cmpc> i think he glib-ified everything in his branch ... so its likely things like false/true might come from glib
| |
| 11:19 | <vagrantc> ogra_cmpc: he used 1 and 0
| |
| 11:19 | <ogra_cmpc> heh
| |
| 11:19 | even better saves you a header or define completely
| |
| 11:20 | <vagrantc> actually, glib stuff was pulled out
| |
| 11:20 | ogra_cmpc: and is totally unreadable, in my opinion
| |
| 11:20 | <ogra_cmpc> oh, why is that ?
| |
| 11:20 | <vagrantc> well, glib includes were pulled out
| |
| 11:20 | abadger1999 has quit IRC | |
| 11:21 | slashdot1x has joined #ltsp | |
| 11:21 | <ogra_cmpc> but glib cares for all the security issues, its way easier to program without introducing buffer overflows or leaks with glib
| |
| 11:22 | and it has all teh nifty utf8 handling already built in so translating is a charm
| |
| 11:22 | <vagrantc> with the amount of shell programming i'm doing, defining TRUE/FALSE and using those in the code is way more likely to not create broken code than having to mentally think through what is true and what is false.
| |
| 11:22 | if there's something else that is not 1 and 0, i'm all for it.
| |
| 11:22 | but give me anything other than 1 and 0 to define true/false.
| |
| 11:23 | <ogra_cmpc> well, its two lines per .c file to add TRUE/FALSE
| |
| 11:23 | then you cqan use both ;)
| |
| 11:23 | <vagrantc> two very worthwhile lines :P
| |
| 11:24 | slashdot1x has left #ltsp | |
| 11:30 | * vagrantc tries re-introducing the includes one at a time | |
| 11:30 | <vagrantc> glib-object.h
| |
| 11:35 | and i foudn the other error ...
| |
| 11:35 | elisboa has quit IRC | |
| 11:35 | elisboa has joined #ltsp | |
| 11:35 | <vagrantc> pushing changes now ...
| |
| 11:40 | joebake1 has joined #ltsp | |
| 11:42 | <warren> How did it build without that inclue?
| |
| 11:42 | <vagrantc> it didn't
| |
| 11:43 | two things happend ...
| |
| 11:43 | two unrelated commits i made broke in different ways
| |
| 11:44 | i had thoguht you were referring to the first one when i said i tested it ... hwoever, i had only tested the branch, i didn't actually test that my merge was legit ...
| |
| 11:44 | i never tested the TRUE/FALSE related patches
| |
| 11:44 | abadger1999 has joined #ltsp | |
| 11:44 | <vagrantc> sorry about all that.
| |
| 11:46 | urg.
| |
| 11:46 | pascal didn't really understand what i meant when i said only the parts of the log that were relvent...
| |
| 11:48 | warren: working for you now?
| |
| 11:49 | mccann has joined #ltsp | |
| 12:02 | <warren> vagrantc, yes, builds now, thanks.
| |
| 12:03 | <vagrantc> warren: again, sorry about all that. i will be more thorough in the future.
| |
| 12:03 | <warren> vagrantc, np, I'm very appreciative of your work.
| |
| 12:06 | <ogra_cmpc> vagrantc, as long as you dont do it on a released branch in the future :)
| |
| 12:06 | trunk is in flux anyway ...
| |
| 12:06 | so breakage can always happen
| |
| 12:06 | * warren tests the current ldm-trunk | |
| 12:11 | * vagrantc tests it too | |
| 12:16 | tux_440volt has joined #ltsp | |
| 12:19 | milesd has joined #ltsp | |
| 12:24 | <warren> wow, directx is so much faster
| |
| 12:24 | * warren hasn't used it in a while. | |
| 12:34 | elisboa has quit IRC | |
| 12:35 | elisboa has joined #ltsp | |
| 12:37 | <xcasex> vagrantc: btw, problem solved after using your packages btw
| |
| 12:40 | <vagrantc> xcasex: cool. they're in debian sid now.
| |
| 12:41 | <xcasex> excellent :)
| |
| 12:43 | <warren> vagrantc, what package normally installs the rc.d directory?
| |
| 12:43 | * warren is missing it entirely | |
| 12:44 | <ogra_cmpc> ldm should
| |
| 12:45 | <vagrantc> no it doesn't anymore
| |
| 12:45 | <ogra_cmpc> (not sure it does though, i never install it without ltspfs which uses teh dir as well)
| |
| 12:45 | vagrantc, well, it should :) even if the dir is empty
| |
| 12:45 | <warren> Oh I see, it creates an empty directory
| |
| 12:45 | <vagrantc> when we moved the ltspfs specific plugins to ltspfs, then those create it
| |
| 12:45 | ogra_cmpc: will result in extra filesystem calls if ltspfs isn't installed
| |
| 12:46 | <ogra_cmpc> hrm
| |
| 12:46 | <vagrantc> but if scotty's xauth plugin works, then it'll be included and that's that.
| |
| 12:46 | <warren> vagrantc, using scotie's ldm xauth solution rc.d would be needed
| |
| 12:46 | elisboa has quit IRC | |
| 12:46 | <ogra_cmpc> yeah, its cleaner to have it there
| |
| 12:47 | * vagrantc doesn't see the cleanliness of having unused directories installed | |
| 12:47 | <ogra_cmpc> locate ldm would return it for example
| |
| 12:47 | <vagrantc> warren: and when we use scotty's xauth code we'll include it
| |
| 12:47 | <ogra_cmpc> it makes it easier for admins
| |
| 12:48 | <warren> when are hte rc.d scripts run?
| |
| 12:48 | concurrently with ldm greeter?
| |
| 12:48 | or upon ssh login?
| |
| 12:48 | <vagrantc> warren: at login time
| |
| 12:48 | <ogra_cmpc> sfter the session is estab;lished
| |
| 12:48 | <warren> isn't that racy then?
| |
| 12:48 | <vagrantc> warren: it's very linear
| |
| 12:48 | <ogra_cmpc> ther are two ssh calls atm
| |
| 12:48 | <warren> "Hey server, send the X display to this address."
| |
| 12:48 | "oops, I'm not allowing you yet."
| |
| 12:48 | "now I am"
| |
| 12:48 | <vagrantc> oh
| |
| 12:48 | that
| |
| 12:48 | ElDios has quit IRC | |
| 12:49 | <ogra_cmpc> the first one does the ground work and gets the socket and tunnel up, the second starts the X session
| |
| 12:49 | * warren reads | |
| 12:49 | <ogra_cmpc> rc scripts are executed before the second one
| |
| 12:49 | <vagrantc> ogra_cmpc: is the second called before or after rc.d ?
| |
| 12:49 | <warren> ok
| |
| 12:49 | <ogra_cmpc> at least it was like that in the beginning
| |
| 12:49 | when they were new
| |
| 12:50 | elisboa has joined #ltsp | |
| 12:50 | * ogra_cmpc wonders how one is supposed to have CK authentication with a regular startx session ... | |
| 12:51 | Guaraldo has quit IRC | |
| 12:51 | <ogra_cmpc> warren, do you know if fedora has a fix for that ?
| |
| 12:51 | (not talking about ltsp startx here)
| |
| 12:51 | Guaraldo has joined #ltsp | |
| 12:51 | <ogra_cmpc> apparently CK only works with display managers ...
| |
| 12:52 | so running startx manually leaves you screwed
| |
| 12:54 | <warren> ogra_cmpc, AFAIK that's still broken here
| |
| 12:54 | <ogra_cmpc> ah
| |
| 12:54 | many users will complain
| |
| 12:55 | <warren> I know =(
| |
| 12:55 | <ogra_cmpc> (i dont se a fix at the horizon, so its likely to stay like that in ubuntu)
| |
| 12:55 | oh, and you even release earlier
| |
| 12:55 | <warren> I know =(
| |
| 12:55 | this pressure is pretty bad
| |
| 12:55 | I'm spending more time fixing regular distro things than LTSP
| |
| 12:56 | <ogra_cmpc> you get used to it over time :)
| |
| 12:56 | * ogra_cmpc works on that schedule since three years now | |
| 13:02 | DonSilver has joined #ltsp | |
| 13:04 | <warren> ogra_cmpc, without CK authentication what drawbacks do you have?
| |
| 13:10 | <ogra_cmpc> none of the admin tools work
| |
| 13:11 | <warren> oh
| |
| 13:11 | =(
| |
| 13:11 | ogra_cmpc, but remote sound and ltspfs still works?
| |
| 13:14 | spectra has joined #ltsp | |
| 13:18 | abadger1999 has quit IRC | |
| 13:20 | <ogra_cmpc> indeed
| |
| 13:20 | both isnt bound to CK
| |
| 13:20 | the gnome tools are lately
| |
| 13:20 | <warren> Oh, I see.
| |
| 13:21 | <ogra_cmpc> they work with sudo though from the commandline
| |
| 13:21 | but thats rather suboptimal
| |
| 13:28 | tux_440volt has quit IRC | |
| 13:33 | tux_440volt has joined #ltsp | |
| 13:36 | <warren> ogra_cmpc, I was just told about a patch for openssh for CK from Ubuntu
| |
| 13:36 | ogra_cmpc, does Ubuntu already ship that patch?
| |
| 13:37 | <ogra_cmpc> yes, but it doesnt solve the prob ...
| |
| 13:37 | non local sessions dont get admin rights at all
| |
| 13:38 | tux_440volt has quit IRC | |
| 13:38 | <ogra_cmpc> i will look again into that over easter and hope to get it fioxed, i'll mail you the patches then
| |
| 13:41 | <vagrantc> ogra_cmpc: this CK stuff is in hardy already?
| |
| 13:42 | <ogra_cmpc> sadly, yes
| |
| 13:42 | its immature imho ... but the upstream tools all switched over
| |
| 13:42 | slipttees has joined #ltsp | |
| 13:42 | <ogra_cmpc> so no choice
| |
| 13:43 | <slipttees> ltsp 5 crap ltsp 4.2 forever
| |
| 13:43 | <vagrantc> ok, now we're only waiting on hppa and mips builds for ldm/ltsp :)
| |
| 13:44 | <laga> slipttees: maybe we can arrange a refund?
| |
| 13:44 | * Guaraldo prepared a new LTSP5 server | |
| 13:44 | <laga> contact the sales dpt
| |
| 13:44 | Pascal_1 has joined #ltsp | |
| 13:44 | * vagrantc posts a "Do not feed the troll" sign | |
| 13:44 | <Guaraldo> vagrantc: How can I change the LDM theme (preparing a custom theme)?
| |
| 13:45 | <laga> vagrantc: not even with sarcasm? :/
| |
| 13:46 | <vagrantc> Guaraldo: use update-alternatives in the chroot ...
| |
| 13:46 | Guaraldo: chroot /opt/ltsp/i386 update-alternatives --config ldm-theme
| |
| 13:46 | Guaraldo: and install your theme into /opt/ltsp/i386/usr/share/ldm/themes/FOO
| |
| 13:46 | <slipttees> laga: it funny!
| |
| 13:47 | <Guaraldo> vagrantc: thanks
| |
| 13:47 | <vagrantc> Guaraldo: with the latest LDM, you can just specify LDM_THEME in lts.conf
| |
| 13:47 | but i don't think that's in gutsy ... not sure about hardy
| |
| 13:47 | <slipttees> laga: why ltsp 5 automatically ?
| |
| 13:48 | dtrask has quit IRC | |
| 13:49 | <Guaraldo> I'm using Gutsy
| |
| 13:49 | LTSP5 server structured today!
| |
| 13:49 | <vagrantc> cool :)
| |
| 13:50 | * warren doesn't see where rc.d happens between two ssh commands | |
| 13:52 | <warren> oh
| |
| 13:52 | I'm blind
| |
| 13:54 | <Guaraldo> vagrantc: My LDM theme doesn't apears on update-alternatives --config ldm-theme
| |
| 13:54 | <vagrantc> Guaraldo: oh, yes ... you'll have to use update-alternatives --install ...
| |
| 13:55 | Guaraldo: look in $chroot/var/lib/dpkg/info/ldm.postinst for an example of how to use it
| |
| 13:57 | joebake1 has quit IRC | |
| 13:57 | joebake1 has joined #ltsp | |
| 13:58 | <Guaraldo> vagrantc: Do I need to rebuild image???
| |
| 13:58 | <vagrantc> Guaraldo: any time you modify anything in the chroot you need to rebuild it, yes.
| |
| 14:08 | elisboa has quit IRC | |
| 14:12 | Pascal_1 has quit IRC | |
| 14:13 | <warren> sbalneav, around?
| |
| 14:24 | vagrantc, do you have hexdump installed?
| |
| 14:24 | /usr/bin/hexdump
| |
| 14:24 | it is from our util-linux-ng which I"m assuming is also in util-linux
| |
| 14:25 | <vagrantc> it's installed
| |
| 14:25 | part of debian's bsdmainutils package ... which is *usually* installed on debian, but it's possible to have an install without it.
| |
| 14:25 | <warren> ok
| |
| 14:25 | <vagrantc> warren: what do you need it for?
| |
| 14:25 | <warren> I figured out how to create a proper xauthority file before X
| |
| 14:25 | hexdump is part of it
| |
| 14:26 | <vagrantc> ah, ok.
| |
| 14:26 | <warren> i'll test this out and see how it goes
| |
| 14:26 | <vagrantc> it'd be nice to get that fixed properly.
| |
| 14:26 | <warren> would be nice if there were a way to launch a xterm *on the client itself*
| |
| 14:27 | <Guaraldo> vagrantc: It worked... My boss has got very impressed... :-D
| |
| 14:27 | vagrantc: Thanks...
| |
| 14:27 | <vagrantc> i've often thought it would be nice to launch a local xterm when doing ldm debugging, rather than having to set up a screen script for it
| |
| 14:27 | Guaraldo: of course! :)
| |
| 14:28 | "LTSP: we'll impress your boss"
| |
| 14:28 | it's practically our motto.
| |
| 14:28 | <Guaraldo> vagrantc: Nice slogan!
| |
| 14:28 | vagrantc: Ops... Nice motto!
| |
| 14:28 | <vagrantc> heh
| |
| 14:29 | <warren> xserverauthfile=$HOME/.serverauth.$$
| |
| 14:29 | what does the $$ do?
| |
| 14:29 | <vagrantc> it's the current PID
| |
| 14:30 | <warren> ah
| |
| 14:30 | <Guaraldo> vagrantc: on the lower right corner of the LDM greeter must be the server name or the station name?
| |
| 14:30 | <vagrantc> Guaraldo: the thin client name
| |
| 14:31 | <Guaraldo> vagrantc: From where does it take it?
| |
| 14:31 | DNS?
| |
| 14:31 | <vagrantc> Guaraldo: if available. if not, it may just set it to "ltsp"
| |
| 14:31 | <warren> ok, initial implementation of this will NOT work with more than one X server
| |
| 14:33 | <vagrantc> i think we can live with that.
| |
| 14:33 | especially if we do something like: /var/run/ldm/serverauth ...
| |
| 14:34 | well ... server_vtN_authfile ...
| |
| 14:34 | but just get *something* working before we hash out the details :)
| |
| 14:34 | <warren> yes
| |
| 14:34 | <vagrantc> and thanks for giving it a go :)
| |
| 14:36 | <warren> no idea if this will work
| |
| 14:36 | i'm confused between server and client auth files
| |
| 14:36 | <vagrantc> yeah, that confused the hell out of me
| |
| 14:36 | <warren> there is a difference?
| |
| 14:36 | =(
| |
| 14:37 | <vagrantc> heck if i know
| |
| 14:37 | * vagrantc is still confused, or this stuff would've been fixed already | |
| 14:38 | <warren> one part of it is:
| |
| 14:38 | mcookie=`dd if=/dev/random bs=16 count=1 2>/dev/null | hexdump -e \\"%08x\\"`
| |
| 14:38 | we used to have a /usr/bin/mcookie binary to do this
| |
| 14:38 | <vagrantc> used to ?
| |
| 14:39 | <warren> F8 had /usr/bin/mcookie
| |
| 14:39 | F9 uses the above
| |
| 14:40 | <vagrantc> eeyk.
| |
| 14:40 | warren: do you know why F9 dropped it?
| |
| 14:41 | <warren> vagrantc, one less binary is a good thing
| |
| 14:41 | vagrantc, looks like it was dropped in upstream X
| |
| 14:41 | we just copied it
| |
| 14:41 | <vagrantc> ah.
| |
| 14:42 | so now, instead of calling mcookie ... we just duplicate that code everywhere ?
| |
| 14:43 | <warren> vagrantc, OTOH it has the same functionality based upon existing components
| |
| 14:43 | abadger1999 has joined #ltsp | |
| 14:43 | <warren> vagrantc, I wonder, TeacherTool or something that allows the teacher to spy on user sessions
| |
| 14:44 | vagrantc, how the heck did that work?
| |
| 14:44 | vagrantc, did it rely on -ac? =)
| |
| 14:46 | <vagrantc> i'm having some weird behavior with ldm-trunk ...
| |
| 14:46 | sometimes it accepts my password and logs in ... and sometimes it takes the password, logs in remotely, and then keeps respawning X for a bit.
| |
| 14:47 | and it's probably related to my patches ....
| |
| 14:47 | better clean up my own messes
| |
| 14:52 | <ogra_cmpc> warren, italc
| |
| 14:52 | <warren> ?
| |
| 14:52 | <ogra_cmpc> http://italc.sourceforge.net/
| |
| 14:53 | italc runs a vnc server as the user ... and uses ssl to authenticate connections from the master
| |
| 14:54 | warren, it was recently fixed to work with ltsp5 (thanks to stgraber who did most of that directly with upstream)
| |
| 14:55 | <warren> ogra_cmpc, ok, that makes sense now.
| |
| 14:55 | glad to hear something isn't taking advantage of -ac =)
| |
| 14:55 | <ogra_cmpc> while i still dont like the ui its the by far most functional tool for classroom management
| |
| 15:01 | abadger1999 has quit IRC | |
| 15:04 | <vagrantc> ah, ldm isn't setting the defaults with "-br -noreset" and such since my patches ... right.
| |
| 15:04 | needs to be set in the screen script, i think.
| |
| 15:04 | or hardcoded...
| |
| 15:08 | <warren> -br -noreset does what?
| |
| 15:08 | -noreset isn't actually doing anything
| |
| 15:08 | or didn't seem to last I tested
| |
| 15:10 | -noreset seems counter productive to me actually
| |
| 15:10 | if it did do anything
| |
| 15:10 | <vagrantc> i think -noreset is what prevents it from doing the crazy stuff i was just seeing
| |
| 15:10 | cliebow_ has quit IRC | |
| 15:10 | <vagrantc> -br just sets the background to black instead of the grey x thatching
| |
| 15:11 | <ogra_cmpc> yeah, dont drop that
| |
| 15:11 | <warren> was -noreset removed from trunk?
| |
| 15:11 | OH!
| |
| 15:11 | <Guaraldo> vagrantc: It's not working...
| |
| 15:11 | <vagrantc> warren: sort of
| |
| 15:11 | <warren> vagrantc, X wont have any clients connected for a split second at first
| |
| 15:11 | vagrantc, which is different from other ways X is used
| |
| 15:11 | <Guaraldo> vagrantc: lower right is only ltsp...
| |
| 15:12 | <warren> -noreset prevents X from killing itself
| |
| 15:12 | <vagrantc> yeah.
| |
| 15:12 | please leave it in.
| |
| 15:12 | <warren> makes sense now
| |
| 15:12 | <Guaraldo> DNS is available...
| |
| 15:13 | <vagrantc> Guaraldo: might be some flawed assumptions in the initramfs-tools script which sets the hostname to ltsp if no other hostname is defined.
| |
| 15:13 | slipttees has quit IRC | |
| 15:13 | <ogra_cmpc> Guaraldo, how about setting them in dhcpd.conf ?
| |
| 15:13 | <vagrantc> Guaraldo: sbalneav said he tested it
| |
| 15:13 | <ogra_cmpc> vagrantc, i.m not sure it does a dns lookup
| |
| 15:13 | but dhcp names should work
| |
| 15:13 | <vagrantc> ogra_cmpc: i'm not sure how it gets the hostname in the initramfs phase
| |
| 15:14 | <Guaraldo> ogra_cmpc: To near 1000 stations?
| |
| 15:14 | <ogra_cmpc> dhcp sends it if it has it
| |
| 15:14 | <vagrantc> Guaraldo: you could also tweak /etc/hostname at boot with a little init script
| |
| 15:15 | <Guaraldo> vagrantc: I'll try...
| |
| 15:15 | <vagrantc> or from the initramfs ...
| |
| 15:15 | slipttees has joined #ltsp | |
| 15:15 | <vagrantc> if you have some predictable ip to hostname mapping ...
| |
| 15:15 | that's what we do at freegeek to avoid setting up dns :)
| |
| 15:16 | <ogra_cmpc> there is nothing that could do a lookup in intramfs by default i think
| |
| 15:17 | <vagrantc> should i just rip out the setting default xserver arguments from within ldm and leave it to the screen script?
| |
| 15:18 | <ogra_cmpc> well
| |
| 15:19 | can we probably store them more centralized ?
| |
| 15:19 | <vagrantc> found another bug, too ...
| |
| 15:19 | <ogra_cmpc> i would love to not hurt my eyes with the default X pattern on startx as well for example
| |
| 15:19 | <vagrantc> well, with the patch i applied ... ldm passes all arguments beyondd the first two to the server
| |
| 15:20 | <ogra_cmpc> things like -br and -noreset should got into a central file we source from the screen script
| |
| 15:20 | <vagrantc> and if ldm has 3+ arguments, it doesn't set the defaults.
| |
| 15:20 | abadger1999 has joined #ltsp | |
| 15:20 | <vagrantc> ogra_cmpc: we can put it in lts.conf, and set if if not defined in the screen script.
| |
| 15:20 | <ogra_cmpc> nah
| |
| 15:20 | <warren> put what?
| |
| 15:20 | <vagrantc> the default xserver arguments
| |
| 15:20 | <ogra_cmpc> i just want something common between the screen scripts
| |
| 15:21 | indradg__ has joined #ltsp | |
| 15:21 | <vagrantc> agreed.
| |
| 15:21 | <ogra_cmpc> no extra lts.conf option
| |
| 15:21 | <vagrantc> well, i don't think we can get rid of it entirely.
| |
| 15:21 | bbut i do believe we can make some common code sourced by all screen scripts
| |
| 15:21 | <warren> -br and -noreset should always be used
| |
| 15:21 | ldm wont work without -noreset
| |
| 15:22 | <vagrantc> warren: even with xdmcp ?
| |
| 15:22 | <warren> oh
| |
| 15:22 | <vagrantc> even with rdesktop ?
| |
| 15:22 | <ogra_cmpc> and br is just friendly to our users
| |
| 15:22 | <warren> with xdmcp you want -terminate instead of -no-reset
| |
| 15:22 | <vagrantc> indeed.
| |
| 15:22 | <warren> rdesktop I don't know how it works yet
| |
| 15:22 | <ogra_cmpc> hmm, so only -br is common yet
| |
| 15:22 | <vagrantc> right...
| |
| 15:23 | <ogra_cmpc> well, that doesnt justifythe effort
| |
| 15:23 | <vagrantc> what effort?
| |
| 15:23 | <ogra_cmpc> to have it separated
| |
| 15:23 | <vagrantc> yes, but -depth and -fp (i know you love it) and some of the others are possibly common as well.
| |
| 15:23 | <ogra_cmpc> i tought we had some more common options to chare
| |
| 15:23 | <vagrantc> but need to be set conditionally.
| |
| 15:24 | and i think we should handle X_CONF from there
| |
| 15:24 | <ogra_cmpc> well, did you ever look ate configure-x.sh ?
| |
| 15:24 | XFT_SERVER is directed to /dev/null
| |
| 15:24 | <vagrantc> not only have i looked at it, i've patched it :P
| |
| 15:24 | <ogra_cmpc> (scotts change, not mine )
| |
| 15:24 | but i fully agree with the decision
| |
| 15:25 | xft simply makes it obsolete
| |
| 15:25 | <vagrantc> i want to support arbitrary X server options and implement common code across all screen scripts, and exactly which options are common i don't really care.
| |
| 15:25 | but there's more than 1.
| |
| 15:25 | <ogra_cmpc> good
| |
| 15:25 | <vagrantc> and it's a trivial amount of code.
| |
| 15:25 | and i will do it.
| |
| 15:26 | in the meantime, i will just fix ldm-trunk to work.
| |
| 15:32 | well, we could just keep "-br" hard-coded ...
| |
| 15:32 | i guess -noreset too ...
| |
| 15:38 | slipttees has quit IRC | |
| 15:40 | indradg_ has quit IRC | |
| 15:40 | <vagrantc> ok, fixed a bug in the screen script and just went ahead and always add -br and -noreset ... (and -ac with LDM_DIRECTX)
| |
| 15:41 | this time i at least tested that it built :)
| |
| 15:42 | <ogra_cmpc> so the options are now all in the screen script ?
| |
| 15:42 | <vagrantc> no
| |
| 15:43 | <ogra_cmpc> hmm
| |
| 15:43 | <vagrantc> "-br" and "-noreset" are hard-coded.
| |
| 15:43 | and "-ac" *cringe*
| |
| 15:43 | but other options all are
| |
| 15:43 | <ogra_cmpc> i wonder what X says if it gets i.e. -noreset from the commandline as well
| |
| 15:43 | i.e. if an admion plays with the options and doesnt know about the internal valuies we set
| |
| 15:44 | <warren> it uses the last one
| |
| 15:44 | -reset is default
| |
| 15:44 | <ogra_cmpc> i would expect it to handle that but who knows, its X
| |
| 15:44 | <vagrantc> i think it uses the last one, but i'll experiment
| |
| 15:44 | <warren> -noreset and -terminate are other options
| |
| 15:44 | If ldm were written properly we wouldn't need -noreset
| |
| 15:44 | <ogra_cmpc> bah
| |
| 15:45 | <vagrantc> all the more reason to keep it in ldm.c, then
| |
| 15:45 | <warren> keep -noreset and -br in ldm
| |
| 15:45 | until ldm doesn't need -noreset anymore
| |
| 15:45 | <ogra_cmpc> warren, its not worse than gdm
| |
| 15:45 | <warren> but I see no reason to want anything else than -br
| |
| 15:45 | <vagrantc> warren: already done.
| |
| 15:46 | Guaraldo has left #ltsp | |
| 15:59 | indradg_ has joined #ltsp | |
| 16:04 | <warren> ok, my xauth attempt didn't work
| |
| 16:04 | i'm seriously confused
| |
| 16:04 | <vagrantc> that's what xauth is for
| |
| 16:04 | if it was clear an straightforward it would be easy to get haXz0r3d
| |
| 16:05 | <warren> you need both a client and server side xauth file
| |
| 16:05 | <vagrantc> and they can't be the same?
| |
| 16:05 | <warren> well
| |
| 16:06 | <vagrantc> and of course, we add the wonders of client/server terminology confusion to the mix ...
| |
| 16:09 | <ltsppbot> "vagrantc" pasted "dpkg-shlibdeps issues" (11 lines) at http://pastebot.ltsp.org/481
| |
| 16:09 | <vagrantc> any ideas what those messages are about ?
| |
| 16:09 | i get them whenever building ldm on sid ...
| |
| 16:10 | milesd has quit IRC | |
| 16:11 | <xcasex> this is probably stupid. all nub questions are. where do i flick the switch for enabling gdm instead of ldm?
| |
| 16:11 | <vagrantc> xcasex: you want to use gdm on the server, or locally ?
| |
| 16:11 | <xcasex> on the server.
| |
| 16:11 | <vagrantc> xcasex: install gdm on the server and configure it to do XDMCP ...
| |
| 16:11 | * warren wonders what scott meant by /tmp/foople | |
| 16:11 | <xcasex> that straightforward?
| |
| 16:12 | <warren> OH!
| |
| 16:12 | <vagrantc> xcasex: set SCREEN_07=xdmcp
| |
| 16:12 | <warren> /tmp/foople is only logging
| |
| 16:12 | <vagrantc> xcasex: why do you want to use gdm ?
| |
| 16:12 | <xcasex> vagrantc: remind me to get you beer somehow :)
| |
| 16:12 | vagrantc: technical or aesthetic reason?
| |
| 16:12 | <vagrantc> xcasex: if it's for performance reasons ... you could probably use LDM_DIRECTX
| |
| 16:12 | <xcasex> :D
| |
| 16:12 | <vagrantc> xcasex: well, using xdmcp means that local devices and remote sound won't work out of the box
| |
| 16:12 | <xcasex> i have already flipped that :)
| |
| 16:12 | ugh
| |
| 16:13 | theme hacking it is then
| |
| 16:13 | <vagrantc> heh
| |
| 16:13 | <warren> PROTOCOL=$(xauth list | awk '{print $2}')
| |
| 16:13 | KEY=$(xauth list | awk '{print $3}')
| |
| 16:13 | this part depends on proper xauth config before X starts
| |
| 16:13 | which I still don't get
| |
| 16:13 | <vagrantc> warren: remember what i said about it depending on configuring xauth ?
| |
| 16:13 | :P
| |
| 16:13 | Gadi has left #ltsp | |
| 16:14 | <vagrantc> sbalneav: we need your xauth wisdom! :)
| |
| 16:15 | * vagrantc feels like a fish flopping on the bottom of a baot | |
| 16:15 | <vagrantc> s,baot,boat,
| |
| 16:15 | <warren> I'm getting close
| |
| 16:20 | <vagrantc> do we need to source ltsp_config in our screen scripts? on debian, they're always called by something which has already sourced them, and all of the variables get exported
| |
| 16:20 | indradg__ has quit IRC | |
| 16:21 | <vagrantc> ah, we could set a variable in ltsp_config ... LTSP_CONFIGURED=true ...
| |
| 16:23 | or LTSP_CONFIG_DONE=true
| |
| 16:24 | <warren> OK, if you logged into user test, then /home/test/.Xauthority contains the client auth cookies put there by ssh
| |
| 16:24 | We have to do the equivalent of this
| |
| 16:25 | <vagrantc> warren: to be able to run the greeter ?
| |
| 16:25 | <warren> vagrantc, no
| |
| 16:25 | vagrantc, well
| |
| 16:25 | * warren goes to ask the gdm guys how it works for them | |
| 16:29 | <vagrantc> so ldm, xdmcp and rdesktop can all share: -config, -depth, that option that ogra always complains about, and -br
| |
| 16:29 | as well as X_RAMPERC
| |
| 16:30 | i think that's definitely enough to warrent common code
| |
| 16:31 | warrant
| |
| 16:31 | <mhterres> Hey guys
| |
| 16:31 | I'm receiving the message "this workstation isn't authorized to connect"
| |
| 16:32 | I know this is a question about lstp-update-sshkeys
| |
| 16:32 | I ran ltsp-update-sshkeys and then ltsp-update-image
| |
| 16:32 | but the problem persist
| |
| 16:32 | any ideas ?
| |
| 16:32 | :-)
| |
| 16:55 | <warren> what is X_RAMPERC?
| |
| 16:55 | <vagrantc> warren: it's the ulimit thing to limit how much memory the X server can take, i think.
| |
| 16:55 | warren: i don't fully understand it
| |
| 16:56 | <warren> vagrantc, talked with the gdm guys, I understand it a little better, they suggested copying the way gdm does it with uses libXau directly instead of running xauth add, sorta makes sense.
| |
| 16:56 | <xcasex> i bet compiz on a via based thinclient is a no go?
| |
| 16:56 | <warren> vagrantc, cleaner code to have it done within ldm
| |
| 16:56 | vagrantc, but this will take a lot more work
| |
| 16:56 | <xcasex> man my sentences dont even come out right when i'm this tired.
| |
| 16:56 | "i bet compiz doesnt work with a via based graphics chipset on a thin client"
| |
| 16:56 | is what i meant.
| |
| 16:56 | <warren> xcasex, does any 3D work on via at all?
| |
| 16:57 | AFAIK no
| |
| 16:57 | <xcasex> precisely
| |
| 16:57 | luckily i can switch it out :>
| |
| 16:59 | <warren> vagrantc, oh interesting things I learned too
| |
| 16:59 | vagrantc, -ac alone doesn't make it secure
| |
| 17:00 | mhterres has quit IRC | |
| 17:00 | <warren> vagrantc, in the case of LDM_DIRECTX=false you will want "-nolisten tcp" so there's no chance anything else on the network can screw with X, because otherwise it is still open.
| |
| 17:06 | <vagrantc> warren: i'm pretty sure that will entirely block all requests from the server as well
| |
| 17:06 | warren: as that's all done with tcp connections
| |
| 17:06 | <warren> vagrantc, oh?
| |
| 17:06 | m
| |
| 17:07 | Oh, I see.
| |
| 17:07 | mccann has quit IRC | |
| 17:07 | <warren> vagrantc, oh, I remember now,
| |
| 17:07 | vagrantc, -ac still allows anybody from localhost to connect
| |
| 17:08 | vagrantc, which is the entire reason why ssh -X still works now
| |
| 17:08 | if you have no xauth cookies
| |
| 17:08 | <vagrantc> -ac allows anybody from anywhere to connect
| |
| 17:08 | abadger1999 has quit IRC | |
| 17:08 | <warren> I mean
| |
| 17:08 | by removing -ac
| |
| 17:08 | <vagrantc> ah.
| |
| 17:08 | <warren> removing -ac but having no xauth cookies
| |
| 17:09 | <vagrantc> which isn't so bad in most LTSP environemnts ... but needs to be fixed
| |
| 17:09 | <warren> yeah, we need to fix the lack of xauth cookies on the X server anyway
| |
| 17:10 | <vagrantc> well, i just committed a patch which should give us a few less filesystem calls ...
| |
| 17:11 | <warren> umm
| |
| 17:11 | what if you don't want to specify a depth at all?
| |
| 17:11 | because you want X to decide itself
| |
| 17:11 | <vagrantc> then don't define depth
| |
| 17:11 | or don't define X_COLOR_DEPTH
| |
| 17:11 | <warren> oh
| |
| 17:11 | <vagrantc> but that's not the patches i just committed
| |
| 17:12 | <warren> misunderstood
| |
| 17:25 | vagrantc, too burnt out for today, leaving office now
| |
| 17:26 | vagrantc, I think a medium-term thing would be to use libXau directly like gdm, but in the short-term I think the shell script method is faster. I'm going to next try to setup xauth MANUALLY from the command line only to learn how it works.
| |
| 17:26 | vagrantc, ttyl
| |
| 17:26 | <vagrantc> warren: cool. then you should have a firm grasp of it :)
| |
| 17:28 | * vagrantc chuckles every time ltsp gets an upload for hurd | |
| 17:28 | <vagrantc> i cannot fathom it actually works
| |
| 17:29 | slidesinger has quit IRC | |
| 17:49 | vagrantc has quit IRC | |
| 17:51 | DonSilver has quit IRC | |
| 18:01 | K_O-Gnom has joined #ltsp | |
| 18:09 | <loather-work> hurd? does it even boot?
| |
| 18:26 | moa_ has joined #LTSP | |
| 18:32 | <ogra_cmpc> it will ...net year, when its ready
| |
| 18:32 | *next
| |
| 18:32 | :)
| |
| 18:33 | <johnny> poor HURD..
| |
| 18:36 | vagrantc has joined #ltsp | |
| 18:39 | <loather-work> poor HURD? poor anyone stuck using it :(
| |
| 18:44 | dtrask has joined #ltsp | |
| 18:45 | dtrask has joined #ltsp | |
| 18:50 | dtrask has quit IRC | |
| 18:52 | staffencasa_ has joined #ltsp | |
| 18:56 | dtrask has joined #ltsp | |
| 18:57 | staffencasa_ has quit IRC | |
| 19:00 | dtrask has quit IRC | |
| 19:06 | elisboa has joined #ltsp | |
| 19:06 | staffencasa has quit IRC | |
| 19:13 | <moquist> stgraber: ping
| |
| 19:13 | mischko has joined #ltsp | |
| 19:16 | <vagrantc> anyone in new york got a couch to crash on on the 29th ?
| |
| 19:22 | <mnemoc> .oO
| |
| 19:46 | K_O-Gnom has quit IRC | |
| 19:49 | <moa_> ok, ok I need some help with sound.
| |
| 19:49 | anyone in here not sleeping?
| |
| 19:50 | K_O-Gnom has joined #ltsp | |
| 19:55 | TelnetManta has joined #ltsp | |
| 19:58 | <vagrantc> moa_: what linux distro and release?
| |
| 19:58 | <MacIver> moa_: what kind of help?
| |
| 19:59 | <moa_> hey, you guys aren't sleeping!
| |
| 19:59 | gentoo 2007.0, ltsp4.2 ( i know, I need to update )
| |
| 19:59 | <johnny> there is no ltsp5 for gentoo, yet
| |
| 20:00 | i'm working on it
| |
| 20:00 | <vagrantc> johnny and dberkholz have been working on ltsp5
| |
| 20:00 | <johnny> just running into bits of problems here and there
| |
| 20:00 | stuck on the initramfs for the moment
| |
| 20:01 | * MacIver gets johnny some tar remover | |
| 20:01 | <johnny> it ain't tar.. it's cpio :)
| |
| 20:01 | hehe
| |
| 20:01 | <MacIver> :-D
| |
| 20:01 | <moa_> Any tips/tricks for getting sound to work with esd/alsa?
| |
| 20:01 | <MacIver> moa_: pulse rocks ;)
| |
| 20:01 | <johnny> i've never used ltsp4.. so no idea
| |
| 20:02 | <moa_> pulse?
| |
| 20:02 | <johnny> pulseaudio
| |
| 20:02 | <MacIver> i think you might be able to hack it with esd, if you have the patience
| |
| 20:02 | <johnny> i gotta eat dinner tho
| |
| 20:02 | bbiab
| |
| 20:02 | <moa_> I've been hacking esd for the past four hours
| |
| 20:02 | <MacIver> heh
| |
| 20:03 | moa_: teaser...on edubuntu, apt-get install pulseaudio, and pretty much have sound for /free/ ;)
| |
| 20:03 | <moa_> It's very strange. I can push mplayer over esd and it works just fine. But any window manager I try has no luck.
| |
| 20:04 | J45p3r has joined #ltsp | |
| 20:04 | <MacIver> moa_: is the variable set before the wm starts?
| |
| 20:04 | <moa_> I'm pretty sure
| |
| 20:04 | <MacIver> how are you setting it?
| |
| 20:08 | <moa_> I'm not actually setting it manually, but kde for instance detects it in the sound settings.
| |
| 20:10 | and env shows espeaker correctly
| |
| 20:30 | J45p3r has quit IRC | |
| 20:55 | vagrantc has quit IRC | |
| 21:10 | Al-Khouli has joined #ltsp | |
| 21:21 | mischko has quit IRC | |
| 21:26 | K_O-Gnom has quit IRC | |
| 22:01 | OvaKill has joined #ltsp | |
| 22:05 | F-GT has quit IRC | |
| 22:11 | tux_440volt has joined #ltsp | |
| 22:42 | moa_ has quit IRC | |
| 22:49 | vagrantc has joined #ltsp | |
| 22:56 | tux_440volt has quit IRC | |
| 22:59 | <vagrantc> wheeee: bzr get http://llama.freegeek.org/~vagrant/bzr/ltsp/features/ltsp-common-x-screen.d
| |
| 23:43 | spectra has quit IRC | |